A geographic comparison of Plato's Atlantis and Ireland as a test of themegalithic culture hypothesis

U. ErlingssonErlingsson Sub-Aquatic Surveys, Sweden

ABSTRACT

Based on geographic similarities, Erlingsson(2004) presented the scientific hypothesis thatthe Atlantean Empire was modelled on themegalithic culture of Europe and NorthernAfrica. It follows from the hypothesis that theisland of Atlantis must have been Ireland.

This is tested scientifically using twogeographical tests, one regarding length andwidth, the other regarding the plain surroundedby mountains. Each of the tests passes at the 2percent confidence level. In view of thesestatistically significant matches, the hypothesisis retained.

Similarities are also found with Irisharchaeology. Newgrange matches the temple ofPoseidon in several aspects, while Knowthmatches a second temple on Atlantis. In Irishfolklore Newgrange is the mansion of a riverdivinity, just like Atlantis’ main temple is.

The sinking of Atlantis is found to haveparallels with how Dogger Bank sank. The timeof the disaster given by Plato corresponds to anearlier dramatic flooding of a North Sea plain,which has been pointed out as a promising areafor underwater archaeology.

The conclusion is that Plato rearrangedfactual information to create a fictional tale ofAtlantis. The traditions can have been passed ondirectly from the megalithic culture to the OldKingdom of Egypt, as they were contemporary.

Even though the facts are rearranged, theyare unique in providing information from theearliest known advanced culture in Europe,many thousands of years before the start ofrecorded history.

1. INTRODUCTION

The only original source on Atlantis is Plato, inthe dialogues Timaios and Kritias (Timaeus andCritias in Latin). If the tale is true to someextent, it may represent our oldest extanttraditions, at least outside Mesopotamia. If itcan be deciphered, it may thus add invaluableclues to the interpretation of the more silentdiscoveries made in archaeology.

We can safely assume that not everything inPlato’s tale matches reality, if not for otherreasons so for errors in the transmission andtranslation. The name Atlantis in a strict senserefers to Plato’s exact description. However, ina lax sense it may be used to refer to a realworld model for only a part of Plato’s tale, justlike a modern city can function as a backdrop toa novel (and be modified in the novel comparedto reality).

The literature on how to interpret Plato’sAtlantis tale is huge, but often of little scientificsubstance. The signal to noise ratio is very poor,which makes it challenging to discern a researchfrontier. Having said that, it is my understandingthat there does not exist any hypothesisregarding the interpretation of the tale that hasstood up to scientific scrutiny. I therefore askfor your indulgence when I now start a scientificstudy from scratch, building on nothing but thetranslated dialogues of Plato.

1.1 The scientific method

To suggest that a certain place may be a modelfor Atlantis in some aspect is to suggest anhypothesis. As a test of the hypothesis, thestudent compares detail after detail. Any

similarity that the student made use of to comeup with the hypothesis, is banned for use in atest.

If a test nevertheless can be devised usingindependent data, and the result is negative, thescientific method dictates that the hypothesisshould be dismissed. However, since we aredealing with a potentially ancient tradition wemust allow for some errors to have entered thetale, and can therefore not be that strict.

Unfortunately, the possibility of such errorsin the tradition largely defeats the purpose of thehypothetic-deductive method. Clearly someother rejection criterion is needed, lest we bestuck with the present deluge of untestedhypotheses.

The solution is to apply the statisticalsignificance test on the null hypothesis. What itamounts to is to raise the bar for a test to pass,by requiring that it be statistically improbablethat it passes by chance.

Another benefit of using the null hypothesisis that segments of the tale can be evaluatedindependently from each other. It is notnecessary to assume that all or nothing must betrue.

This study is based on the classic scientifichypothetic-deductive method and statisticalsignificance-tests on the null hypothesis. Asignificance level of 0.02 (2%) was chosen asrejection criterion. The null hypothesis is thatPlato made it all up, and that any similarity withreality is purely coincidental.

1.2 Why the case is open

The mainstream opinion is that the Atlantis taleis entirely fictional. However, this has not andcan not be proven.

One flawed line of argument is this: “If it isfictional it exists to serve the story. It serves thestory. Thus it is fictional.” The logical operator(if A so B) is false only when A is true and B isfalse. When, as in this case, B is true, A can beeither false or true. Thus, the argument islogically false (illogical). An argument has to beboth logically true and relevant to be valid, whythis one is invalid.

Christopher (2001) is another example of aflawed argument. Plato advocated the use of liesin the service of propaganda to maintain amilitaristic, hierarchically structured society,

and to prevent real democracy. We cantherefore not rule out that Plato may haveintended to create a political myth, but even ifthat is the case, it does not imply that everythingis fiction.

However, Christopher, a linguist, considersany partial resemblance meaningless. Giventhat, as he states, all legitimate scholars havejettisoned the conclusion that the account isentirely factual and inerrant, it of coursebecomes somewhat of a truism that Atlantisdoes not exist.

Old tales like the Icelandic Sagas werenotorious for mixing fact and fiction. If wereject all old tales as potential sources of factualinformation, we loose a large part of our history.It is by choice, not necessity, that Christophercloses the door to the past.

Any student of prehistory is forced to laypuzzle with incomplete data, and any piece ofinformation may turn out to be crucial. Even ifthe only thing that can be proven from this lineof inquiry is that Plato had access to knowledgefrom a specific time and place, it is still apotentially invaluable result.

Therefore I elect to use a paradigm in whichthe goal is to search for pieces of facts in theAtlantis tale, and to use these pieces to deducehow the traditions were transmitted and the taleassembled by Plato.

All quotes of Plato’s dialogues are fromBenjamin Jowett’s translation unless otherwisestated.

2. HYPOTHESIS

2.1 The empire in Kritias

The hypothesis was formulated by Erlingsson(2004) based on information in Timaios only(not Kritias), notably this passage: “This powercame forth out of the Atlantic Ocean, for inthose days the Atlantic was navigable; and therewas an island situated in front of the straitswhich are by you called the Pillars of Heracles;the island was larger than Libya and Asia puttogether, and was the way to other islands, andfrom these you might pass to the whole of theopposite continent which surrounded the trueocean; for this sea which is within the Straits ofHeracles is only a harbour, having a narrowentrance, but that other is a real sea, and the

surrounding land may be most truly called aboundless continent.”

The text seems to be referring to the NorthAtlantic islands, which can be used to divide thecrossing to America into many short legs, likethe Vikings did. In my experience, it is by farthe best route to take across the ocean for a boatwithout proper sleeping quarters. During theclimatic optimum in the Atlantic Period, it willhave been even easier to make a crossing at thislatitude.

The text in Timaios continues: “Now in thisisland of Atlantis there was a great andwonderful empire which had rule over thewhole island and several others, and over partsof the continent, and, furthermore, the men ofAtlantis had subjected the parts of Libya withinthe columns of Heracles as far as Egypt, and ofEurope as far as Tyrrhenia.”

Megalithic tombs (Fig. 1) are found onseveral islands in the Atlantic Ocean, on parts ofthe European continent, and, furthermore, theparts of Africa inside the Straits of Gibraltaralmost as far as Libya, and in Europe as far asthe Tyrrhenian Sea and southernmost Italy (Fig.2). Considering that “the pillars of Heracles”was understood in Plato’s time to mean ‘theStraits of Gibraltar’, and that “Libya” meant‘Africa’, the extent of the empire of Atlantis inPlato’s account matches that of the megalithictombs quite well.

I therefore hypothesize that the empire ofAtlantis refers to the entity responsible for theerection of the so-called megalithic tombs inEurope and northern Africa, an entity that I forshort call the megalithic culture. I avoid theword tomb since they also erected standing

stones (menhirs, stone circles), and since it is farfrom obvious that the primary function of all so-called tombs were as burial places.

The construction of the megalithic tombsended suddenly around 2,800 BC, in connectionwith a climatic deterioration. Note that the termmegalithic culture in this sense does not implyany connection with megaliths on othercontinents.

2.2 Deduction

In the hypothetic-deductive method a predictionis made from the hypothesis, and tested usingindependent data.

Plato let us understand that the empireAtlantis was based on a large island (after whichthe empire was named) along the path from theMediterranean to the opposite continent(America), and that they completely controlledthat island. A quick look at Figure 2 shows thatthere are only two large islands off Europe, andthat of those, only Ireland is completely withinthe distribution area of megalithic tombs.

I therefore deduce that for the hypothesis tobe correct, Ireland should correspond to thedescription of the island Atlantis.

2.3 Independence criterion

The test must be made using independent data.This is satisfied since the hypothesis and thededuction were formulated without knowledgeof relevant Irish geography, geology,archaeology, or mythology, and furthermorewithout knowing the content of Kritias. Byusing only information from that dialogue, thedata are even doubly independent.

Among the geographical information inKritias is the following: “The whole countrywas said by him to be very lofty and precipitouson the side of the sea, but the countryimmediately about and surrounding the city wasa level plain, itself surrounded by mountainswhich descended towards the sea; it was smoothand even, and of an oblong shape, extending inone direction three thousand stadia, but acrossthe centre inland it was two thousand stadia.”

This information about Atlantis is comparedwith modern geographic data on Ireland,derived from encyclopedias, and from digitalgeographic information system (GIS) databasesin vector and raster format.

Figure 1. The cairn at Haväng in southern Sweden. It is atypical megalithic tomb of the type without passage.

3. DIMENSION TEST

The first test compares the length and width ofAtlantis and Ireland.

3.1 Data

On a digital chart of the world in scale1:250,000, the longest dimension of Ireland wasmeasured to 490.0 km and the greatest width to341.8 km (Fig. 3). This has to be converted tostadia for comparison with Atlantis.

In Greece a stadion was 600 feet, while inEgypt it was 400 Royal cubits. A cubit being 1.5foot, it is the same thing. The Greek andEgyptian stadion had different lengths, though.Since Plato gave all measures with only onesignificant digit, he evidently simply exchangedthe Egyptian units for the corresponding Greekunits.

If Plato’s account is correct, the Egyptiansgot the measures from another country (hisalleged Atlantis). We may assume that also theEgyptians took over the measurement values,regardless of small differences in themeasurement units. Hence it is irrelevant howlong a stadion was in Egypt or Greece. What we

require is the length of a stadion at the time andplace we are interested in: Megalithic Ireland.

Based on the investigation of manymegalithic sites on the British Isles, Thom(1967) found a common denominator of 0.829m. He called it a megalithic yard. Since thereare 200 yards on a stadion, a megalithic stadionequals 166 m. The dimensions of Ireland in thisunit are 2,950 times 2,060 stadia—within 3% ofPlato’s values. Ireland also matches Atlantis byhaving the greatest width “across the centreinland”.

3.2 Test

The real test is, though, if the correspondence isstatistically significant. The null hypothesis isthat Plato invented Atlantis, and happened togive the dimensions of Ireland by chance.

The question thus becomes, what is theprobability that Plato would have stated thedimensions as 3,000 times 2,000 stadia by purechance?

Based on the size of territories in the Antiqueworld, Erlingsson (2004) estimated that areasonable length could have been in the rangeof 700 to 10,000 stadia. Assuming a width of atleast 600 stadia, the number of possible valuesbecomes 91. The probability that the nullhypothesis is correct is thus 0.011, well under0.02, which means that the correspondence isstatistically significant and the null hypothesis isrejected.

An alternative scale-independent calculationcan be made as follows. As the reasonable sizeof the island varies by more than one order ofmagnitude, it would be conservative todisregard the order of magnitude in thedimensions, and only consider the significantdigit. For each of the length and width, the digitcan be any one in the range one through nine.This creates 81 possible combinations and asignificance level of 0.012, why the nullhypothesis still is rejected.

An assumption in both cases is that each sizeor digit has the same chance of being picked. Apoll (www.macpolls.com/?poll_id=332) onlineasked “Pick a number between 1 and 10”.Removing the 10, the number 2 had a frequencyof 4.95%, and 3 had 11.21%, among the 4252respondents. The probability that someonewould pick 2 and 3 can therefore be estimated to

Figure 2. The extent of megalithic tombs and the locationof Newgrange (star). From Erlingsson (2004), which islargely based on Burenhult (1999, Fig. 77).

Figure 3. The elevation of Ireland with the location ofNewgrange (star). The lines show where length and widthmeasurements were taken. Based on ETOPO2, adoptedfrom Erlingsson (2004).

0.0055, which means that the null hypothesis isrejected. (The number that most pick is 7, withalmost one third of the votes.)

3.3 Interpretation

The megalithic yard is still not generallyaccepted. However, the existence of astatistically significant smallest commondenominator in many stone monuments requiresan explanation. The megalithic yard seems to bethe best explanation offered. At any event, it isnot crucial for the statistical significance test, asany value of the stadion between 137 m and 196m would result in the rejection of the nullhypothesis.

Using virtually any other stadion, from theSumerian of 148.5 m to the Olympic of 192.3m, the conclusion holds true. Only the Egyptianstadion is too long with its 209.2 m.

However, as was argued earlier, even if wedo not accept the existence of a megalithicstadion there is still no logical reason to inferthat the Egyptian one was used to measureAtlantis (unless the hypothesis being tested isthat the Old Kingdom of Egypt was a directcontinuation of the Atlantean civilization,which, however, would contradict what Plato

wrote about the war). Furthermore, the Egyptianstadion seems to be an outlier, being about oneseventh longer than the average stadion—thesame relationship as between a common cubitand a Royal cubit in Egypt. There is thus strongreason to suspect that the unit being used,megalithic stadion or otherwise, was less than196 m long.

In evaluating the conclusion, another relevantquestion is if the translation that was usedcorrectly interpreted Plato’s words.

The size and landscape of Atlantis isprimarily described in Kritias 118a. Sometranslations describe the plain as rectangular,but there does not seem to be any tangiblesupport for that in Plato’s text. There are alsosome translations that interpret the dimensions,3,000 x 2,000 stadia, as referring to the plainand not the whole island. Due to the way ofwriting in Plato’s time—without punctuationmarks or even spaces between words—andsince Plato lined up these descriptions withoutrepeating the subject of the sentence (and sincethe measurements are in genitive, which isidentical in every genus), it is not completelyclear what he was referring to.

There is a detail that seems to have beenmissed in translations, all of which give thewidth as two thousand stadia. However, aftermentioning the length, Plato writes “!"#" $%µ%&'( ")' *"+"##,- "(. $/&0/+/.(”, whichI translate as ‘in the middle from the sea morethan two thousand’. Recall that the width ofIreland is 2,060 megalithic stadia—hence morethan two thousand from sea to sea.

4. LANDSCAPE TEST

The second test compares Atlantis’ centralmountain-surrounded plain with Ireland.

4.1 Data

This is what The World Factbook (2004) has tosay about the geomorphology of Ireland:“Mostly level to rolling interior plainsurrounded by rugged hills and low mountains;sea cliffs on west coast”. Compare it with thisquote from Kritias: “The whole island was highand steep on the side of the sea, but at andaround the city the surrounding was a plain,

which in turn was surrounded by mountains thatsloped down to the sea.”

Plato’s text first stresses the sheer cliffs witha plain on top, which also exists on Ireland (Fig.4). Just like The World Factbook, he comple-ments that with mentioning the plain surroundedby mountains.

The Irish “central plain” is clearly visible inFigure 3, which is based on a digital elevationmodel (DEM) of the world, with 2 minutesresolution. The plain is roughly rectangular,extending from the east coast to the mountainson the west coast, between N53º and N54º.

Ireland clearly matches the description asregards the landscape, but the question is againif this similarity is statistically significant, or ifit could have been caused by chance.

4.2 Test

Coastal cliffs are very common on oceanicislands, especially those with a volcanic origin.Mountains and plains are equally common. Buthow frequently is a plain, obviously ofsubstantial dimensions, surrounded bymountains?

Using the ETOPO2 DEM, the landscape ofthe 50 largest islands in the world was evaluatedas regards one specific parameter: If they have aplain surrounded by mountains. All remainingislands were too small to evaluate in a 2 minuteDEM.

A careful scrutiny of the GIS came up withonly one island that had a plain surrounded bymountains, and that is Ireland. As expected, thetypical situation is a central mountain with alow coastland, or a low island withoutmountains.

With only 2% of the islands having a plainsurrounded by mountains the result issignificant, and the null hypothesis is rejected.

The test assumes that Plato would havedescribed Atlantis in a normal and typical wayfor islands if he had been making it up. That is,it assumes that the likelihood for him todescribe a certain feature is directly proportionalto the frequency of that feature on real islands.In the view of this, it is perhaps more relevant tolook at insular geography in Plato’s part of theworld, than on the entire planet.

Such a comparison was made using a DEMwith 1 km resolution. Ireland plus the fivelargest islands in the Mediterranean Sea wereanalyzed as follows: The distance from thecoastline was calculated for each cell, and thecells divided in 1-km distance classes. Theaverage elevation for each distance wascalculated. Both the average elevation and thedistance were normalized.

The 6 resulting curves (Fig. 5) show howclear the distinction is between Ireland on theone hand, and all the major Mediterraneanislands on the other. Ireland alone is lower inthe inland than near the coast.

4.3 Interpretation

A large plain surrounded by mountains violatesthe normal geomorphology of islands, andespecially the situation in Plato’s part of theworld. The description of Atlantis closelyresembles that of Ireland.

5. DISCUSSION

Based on the geographical tests, the hypothesisthat Plato was describing Ireland and themegalithic culture is retained. Before reflectingon how Plato could have access to thisinformation, some other aspects than thegeography will be compared.

5.1 Archaeology

If Ireland is the island Atlantis in Plato’s tale,and the megalithic culture is the empire, thenthe finest megalithic monuments ought to befound on Ireland. A university textbook inarchaeology provides the answer.

Burenhult (1999) writes, “Ireland possessesthe without comparison largest number of

Figure 4. Cliffs of Mohrer on the west coast of Ireland.Composite of two photos ©Mike Goldsman 1997.

megalithic tombs in relation to the surface areaof the island” (p. 284), and “…the choicestmegalithic constructions created in Europe. Tothese belong principally Newgrange, Knowthand Dowth in the Boyne valley, richlyornamented […] The monuments bear witnessto deep astronomical knowledge…” (p. 287; mytranslation). Some of the oldest megalithicmonuments are also found on Ireland, inCarrowmore in County Sligo, dating back to themid sixth millennium BC (Burenhult 1999, p.286; Bergh 1995). Ireland is thus a likely centralregion for the megalithic culture, and thehypothesis holds up.

Newgrange is shown in Figure 6, in a frontalview and a close-up of the entrance. Compare itwith the rather typical megalithic cairn in Figure1, where a rectangular chain of standing stonesmark the sacred area, in the middle of which is atomb created by a few megaliths, with one ofthem serving as roof. It is of similar age asNewgrange.

Plato mentioned that the Atlanteans quarriedwhite, black and red rock, and wrote in Kritias116b, “Some of their buildings were simple, butin others they put together different stones,

varying the colour to please the eye, and to be anatural source of delight.”

The end of it reads in Greek, “µ%/1(2(#%-#'2- + /*'2- )'/!/+" 23"/('( )"/$/"-0"4/(”. According to the online dictionary ofthe Perseus project, µ%/1(2(#%- means ‘mixing’(properly used of liquids), and #'2- +/*'2- is‘the stones’. The last three words mean that theywere ‘weaved together to a childish grace’,literally. The remaining word, )'/!/+", means‘many-coloured, spotted, mottled, pied,dappled’, suggesting that in a façade of stones inone colour, there were spots made of stones in adifferent colour (the form of the word is dualis).

Compare this with the façade of Newgrange(Fig. 6), made up of quartz (white) with spots ofgranite stones (dark). The wall is a restoration,since the original one collapsed over 4,000years ago (the monument was built around3,200 BC; O’Kelly 1982). The decoration isanother match between Atlantis and Ireland. Itmay prove statistically significant since it israther unusual.

Plato mentioned two temples, one being forboth their ancestors: “in the centre was a holytemple dedicated to Cleito and Poseidon, whichremained inaccessible”, while the other was fortheir divine ancestor alone: “Here wasPoseidon's own temple which was a stadium inlength, and half a stadium in width, and of aproportionate height, having a strange barbaricappearance.”

As was argued in Erlingsson (2004), thepassage tomb of Knowth (Eogan 1987) with itstwo chambers may correspond to the temple forboth Cleito and Poseidon, while Newgrangemay correspond to that of Poseidon alone (theIrish monuments are about half a megalithicstadion across). Note that while there arehundreds of burials in Knowth, there may not bea single one in Newgrange. This would seem toagree with the latter being a temple for a god,rather than for a mortal ancestor.

The single chamber in Newgrange has acorbelled vault, whereas in Knowth there is onechamber with a corbelled vault and one with aflat stone slab for roof like in the majority ofmegalithic tombs. When caliph Abdullah Al-Mamun of Baghdad in AD 820 broke into theGreat Pyramid in Egypt, he first found achamber with gabled roof and named it theQueen’s Chamber, since they buried their

0%

20%

40%

60%

80%

100%

0% 20% 40% 60% 80% 100%

Crete

Cyprus

Sicily

Sardinia

Corsica

Ireland

Figure 5. Normalized average elevation plotted againstnormalized cumulative distance from the coastline forIreland and the five biggest Mediterranean islands.

women in tombs with gabled ceilings, but menin tombs with flat ceilings. This distinction

based on sex might be the reason for Knowthhaving two chambers with different roof types,which would be in line with it being a templefor both the earthborn ancestral mother Cleito,and for their divine ancestral father Poseidon.

Plato mentioned that the distance from thesea to the city was 50 stadia (Kritias 115d andpossibly 113c). The Boyne valley monumentsare located 50 stadia from the sea (Erlingsson,2004, p. 50).

In conclusion, there is reason to suspect thatPlato also had access to a description of themonuments in the Boyne valley in CountyMeath, Ireland.

5.2 Mythology

Ireland has a long oral tradition. The location ofNewgrange was unknown during four millennia,after it collapsed into an inconspicuous hillwhen the retaining wall gave way (it is nowconcrete reinforced). In spite of this, Irishfolklore remembered it. The Irish name is Brúna Bóinne, which means ‘Mansion of theBoyne’. The river name Bóinne means ‘whitecow’ (a cognate to Sanskrit Govinda), and it isrelated to the name of the river goddess Boann(or Boand), whose son Oengus now inhabits themansion. Incidentally, the Milky Way is called‘the way of the white cow’ in Irish.

Boann, without permission, approached asacred well. The well sprang up in a flood andtook her out to the sea, thus creating the riverBoyne, a most revered river on Ireland. Thewell belonged to the river god Nechtan, a namecognate to Roman Neptun, i.e., Poseidon,originally a river god. There is thus a connectionbetween Newgrange and Poseidon. Since theword that Plato used for temple also means‘residence of a god’, the Mansion of the Boyneis a namesake of Poseidon’s temple on Atlantis,except for the sex of the river divinity.

Irish mythology contains elements that hintat an ancestral homeland that was flooded bythe rising sea, and myths of sunken cities existon much of the Atlantic seaboard of Europe.However, since such myths are common allover the world they tend to be dismissed.Nevertheless, most of the planet did experiencea significant sea-level rise after the Ice Age, sothe omnipresence of the motif is logical. Thequestion is, though, if the motif has a

Figure 6. Newgrange. Note the granite stones in thequartz façade. The original façade continued straight outto both ends of the decorated entrance stone. Above thedoor there is a light opening, through which the risingsun’s rays reach the grave chamber on Midwinter solstice.

psychological origin common to all people, or ifit is a very ancient tradition.

As a test of whether a myth about geographicchanges can survive for such a long time, wemay look at an island that has experienced theopposite, namely rising from the sea: Gotland inthe Baltic Sea. The creation myth of that island,preserved in the introduction to their law, recallshow the island used to sink in the sea everymorning and rise every evening, until Tjelvarcame and brought fire.

Modern geology has shown that Gotlandmelted out from the inland ice as barren rocksbarely protruding from the Baltic Ice Lake some10,000 years ago. The core element of thetradition is therefore correct, since the mythreflects that this island has risen from the sea,while the majority of coasts have been flooded.This hints at the possibility of very ancientmemories being preserved in oral tradition,especially on remote islands where there hasbeen a long continuous occupation.

5.3 Sinking

Ireland did not sink in the sea, but Atlantis did,according to Plato. For this aspect of the taleIreland is obviously not the model. A distinctpossibility is that it is a memory of some otherplace that was preserved in megalithic Ireland.

Erlingsson (2004, p. 22) suggested a realworld explanation for Atlantis’ sinking, namelythat Dogger Bank was destroyed by theStoregga tsunami. By coincidence, the floodwave came at a time when Dogger Bank wasabout to sink anyway in the rising sea level.

It was a very powerful tsunami (Bondevik etal. 2003) that may well have converted largelow-lying areas into mud banks. This isespecially true if they were protected againststorm floods by coastal dykes or natural sanddunes, since those would prevent the sea fromwithdrawing. The apparent effect would havebeen that the island sank in the sea after adramatic natural disaster. Furthermore, it wouldhave created a mud bank that made the sea un-navigable, just as the case was with Atlantis(Timaios 25c-d and Kritias 108e).

In Timaios 24e Plato positions Atlantisbefore the mouth that the Greeks call thecolumns of Heracles. In this context it is worthnoting Tacitus’ report in Germany and its tribes

(XXXIV) that the columns of Heracles arelocated by Friesland, which in turn is located bythe Rhine. Dogger Bank being located off themouth of the Rhine, one must contemplate thepossibility that it was the Phoenicians who firstapplied the name pillars of Heracles toGibraltar, and that they did so to keep othersaway by exploiting the fear connected with anold myth (Plato apparently considered it wellknown that the sea outside the pillars ofHeracles was un-navigable).

It is also worth noting that at the time whenDogger Bank was sinking it formed a tidal bankthat covered most of the distance between theshores of England and Denmark. It thuseffectively prevented the navigation from theRhine to the open North Sea—just what Platoclaimed happened when Atlantis sank. In starkcontrast, an Aegean island that sinks in anearthquake does not create a mud bank thathinders navigation. As Plato’s description fitsthe geography of Dogger Bank of six millenniaago well, while being so exotic to his own partof the world, the match is noteworthy.

An alternative reason for the apparentsinking of Dogger Bank could be the risingworld sea level after the Ice Age. It is usuallydismissed as a potential cause of disastrousfloods, since the melting of ice is a gradualprocess that should not produce rapidtransgressions. However, some scientists aresuggesting multiple events of rapid sea-levelrise triggered by jökulhlaups and ice-marginallake bursts from the Laurentian inland ice (e.g.,Blanchon & Shaw 1995; cf. Fig. 7). The lastcataclysmic transgression would have drownedDogger Bank, and the previous one would havedrowned a plain to the south of Dogger Bank,considered to be a promising location forunderwater archaeology by Flemming (2002).

This plain, called Doggerland by Coles(1998), would have been flooded in a rathershort time some 9,000 years before Plato. Thisis precisely the number of years Plato said hadpassed since the disaster. That transgressioncoincided with the end of the Ice Age, andrepresented a global change-event on a scalethat makes the gloomiest predictions today seeminconsequential. Incidentally, there was acomparable event of equally rapid sea level andtemperature rise a mere three thousand yearsearlier.

The evidence for these alleged rapidtransgressions are not universally accepted,though; nor are Laurentian jökuhlhaups,although they appear theoretically possiblethrough the captured ice shelf mechanismsuggested by Erlingsson (1994a, b).

What is irrefutable is that Dogger Bankbecame an island during the Holocenetransgression (Isle of Dogger in Fig. 7). Thefinal sinking of this island was the last greatcataclysm before Plato’s (and our) time,whether it was caused by the Storegga tsunamior the final drainage of Lake Agassiz throughHudson Bay.

5.4 The origin of the Atlantis tale

The image that emerges is that megalithicIreland kept alive the memory of the disasterthat struck Dogger Bank, during three thousandyears (Fig. 7). After the collapse of the

megalithic culture, the memories of it, as well asof the Dogger Bank cataclysm, may well havebeen kept in Egypt just as Plato claimed.

As the insert in Figure 7 shows, there is nogap in time between these cultures. Nor is thereany sizeable spatial gap, since the Atlanteanempire, like the megalithic culture, reached theeastern Mediterranean.

It seems a reasonable conclusion that Platogot hold of historic data the way he claimed,and used it to create a political myth in the wayhe himself advocated.

6. CONCLUSION

Plato based the description of Atlantis onmegalithic Ireland, which at the time appears tohave been the focal point of a large culture. Itseems feasible that the memory of the Storeggatsunami on Dogger Bank was kept alive in thatculture, and carried from there to Egypt,eventually to be mixed up with Ireland itself.When Plato wrote his dialogues, he either putthe fragments together in a novel order whileadding some flair, or someone else before himhad already done that.

Although we must take into account thatPlato may have added many fantastic details tothe description of Atlantis in order to glorifyAthens’ victory over the empire, his tale canstill give us clues. Even if the tale is partlyfictional, whatever factual details can beextracted are invaluable, since there simply isno other source of recorded tradition from theEuropean Stone Age.

Incidentally, this result is also proof thatAtlantis in the strict sense never existed.

ACKNOWLEDGEMENTS

The author would like to acknowledge theassistance from friends during the long researchprocess that led up to this study, especially thegenerously offered help from Torgny Frembäckand Athina Tavoultzidou. Thanks are also dueto prof em J.O. Norrman, Å. Sundborg and W.Karlén, whose critique of the analysis has beenmost welcome.

The Perseus project website atwww.perseus.tufts.edu was consulted for theoriginal Greek and Latin texts.

Figure 7. Suggested context of the Atlantis tale in climaticchange and sea level rise. The x-scale is in thousands ofcalendar years before present, defined as 1950 AD. Thethin line is the $18O temperature proxy from Greenland(GRIP Members 1993). The Ice Age ended at 11.5 kBP.The bold line is the sea level from Blanchon and Shaw(1995). There were drastic rises around 11.5 kBP and 7.6kBP. The flood wave symbol marks the tsunami from theStoregga submarine slide around 8.1 kBP. Insert: Thedashed line illustrates the drastic cultural decline ca 7.5kBP from the Kongemose (K) to the Ertebølle (E)Mesolithic culture. Rectangles represent the temporalextent of Dogger Bank as an island, the megalithic culture,Egypt, and Antiquity. The arrows show how the traditionof the sinking may have been transmitted between them.Since there are significant reservoir effects whencalibrating from radiocarbon years to calendar years, it isquite possible that all events around 8 kBP occurred in ashort time span and were related in some way. The earliestmegalithic tombs on Ireland also date from that timeframe.

The paper is dedicated to my father ErlingNilsson, and his father, Knut Nilsson, forproving that anyone can see through propagandaif he only uses his wits and logic.

REFERENCES

Bergh, S. (1995). Landscape of the Monuments. A studyof the passage tombs in the Cúil Irra region, Co. Sligo,Ireland. Arkeologiska undersökningar, Skrifter 6:Riksantikvarieämbetet, Sweden.

Blanchon, P. and Shaw, John (1995). Reef drowningduring the last deglaciation: Evidence for catastrophicsea-level rise and ice-sheet collapse, Geology 23: 4-8.

Bondevik, S., Mangerud, J., Dawson, S., Dawson, A.,Lohne, Ø. (2003). Record-breaking Height for 8000-Year-Old Tsunami in the North Atlantic. Eos 84: 289-300.

Burenhult, Göran (ed.) (1999). Arkeologi i Norden 1.Stockholm: Natur och Kultur.

Christopher, Kevin (2001). Atlantis: No way, no how, nowhere. Skeptical Briefs newsletter, 11(3).www.csicop.org/sb/

Coles, B. J. 1998. Doggerland: a speculative survey.Proceedings of the Prehistoric Society 64: 45-81.

Eogan, George (1987). Knowth and passage-tombs ofIreland (New Aspects of Antiquity). London: Thames& Hudson.

Erlingsson, Ulf (1994a) The ‘Captured Ice Shelf’hypothesis and its applicability to the Weichselianglaciation. Geografiska Annaler 76A: 1-12.

Erlingsson, Ulf (1994b). A computer model along a flow-line of an Ice Dome—‘Captured Ice Shelf’.Geografiska Annaler 76A: 13-24.

Erlingsson, Ulf (2004). Atlantis from a Geographer’sPerspective: Mapping the Fairy Land. Miami:Lindorm Publishing.

Flemming, N. C. (2002). The scope of StrategicEnvironmental Assessment of North Sea areas SEA3and SEA2 in regard to prehistoric archaeologicalremains. UK Department of Trade and Industry,Report Nr TR_014.

GRIP Members (1993). Climate instability during the lastinterglacial period recorded in the GRIP ice core.Nature 364: 203-7.

O’Kelly, Michael J. (1982). Newgrange: Archaeology,Art and Legend (New Aspects of Antiquity). London:Thames & Hudson.

Thom, Alexander (1967). Megalithic Sites in Britain.Oxford: Clarendon Press.

The World Factbook (2004). Washington: CIA.